Berger, W.H., Kroenke, L.W., Mayer, L.A., et al., 1993 Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 130 21. QUATERNARY OXYGEN ISOTOPE RECORD OF PELAGIC FORAMINIFERS: SITE 805, ONTONG JAVA PLATEAU 1 W.H. Berger, 2 T. Bickert, 3 H. Schmidt, 3 G. Wefer, 3 and M. Yasuda 2 ABSTRACT The oxygen isotope records of G. sacculifer and Pulleniatina in the uppermost three cores at Ocean Drilling Program Hole 805C span the last 1.6 m.y., an estimate based on Fourier stratigraphy. The last 700,000 yr are dominated by both eccentricity and obliquity related orbital fluctuations. The range of variation of δ 18 θ values is about 1.5‰, of which ca. 75% may be assigned to global ice volume effect. The remainder of the range is shared by the effects of surface temperature variation, thermocline depth change (in the case of Pulleniatina, especially), and differential dissolution. Before 1 Ma, obliquity related fluctuations dominate. The transition between obliquity and eccentricity dominated time occurs between ca. 1 and 0.7 Ma. It is marked by irregularities in phase relationships, the source of which is not clear. The age of the Brunhes/Matuyama boundary is determined as 794,000 yr by obliquity counting. However, an age of 830,000 yr also is compatible with the counts of both eccentricity and obliquity cycles. In the first case, Stage 19 (which contains the boundary) is coincident with the crest of the 19th obliquity cycle, setting the first crest downcore equal to zero, and counting backward (ol9). In the second, Stage 19 coincides with o20. No evidence was found for fluctuations related to precession (23 and 19 k.y.) rising above the noise level, using plain Fourier expansion on the age model of the entire series. Detailed stratigraphic comparison with the Quaternary record of Hole 806B allows the recognition of major dissolution events (which increase the difference in δ 18 θ values of G. sacculifer at the two sites). These occur at Stages 11 13, 16 17, and near 1.5 Ma (below o33). INTRODUCTION The Ontong Java Plateau straddles the equator and bears a thick cover of calcareous sediments. Already these sediments have contrib uted much information on the stable isotope stratigraphy of the Quaternary (Shackleton and Opdyke, 1973,1976; Schiffelbein, 1984; Hebbeln et al., 1990; Wu and Berger, 1991), as well as to that of the Neogene in general (Woodruff et al., 1981; Whitman and Berger, 1992). Extensive work on box cores (Johnson et al., 1977; Berger et al., 1978, 1987) provides detailed background information on pres ent day conditions of sedimentation, back to the last glacial. During Leg 130 of the Ocean Drilling Program (ODP), advanced hydraulic piston coring (APC) was used to retrieve a number of undisturbed sequences within the Neogene (Fig. 1). We have reported on the stable isotope stratigraphy of the uppermost five cores of Hole 806B elsewhere in this volume (Berger et al., this volume; Bickert et al., this volume; Schmidt et al., this volume). For compari son with this record, we selected the first three cores from Hole 805C (1°13.69'N, 160°31.77'E; water depth, 3188 m) for detailed study of the Quaternary. These cores (130 805C 1H through 3H) comprise the upper 26.8 m of sediment, which consists of calcareous ooze with foraminifers and nannofossils. Preservation is quite good on the whole. The period represented by the sequence sampled is about 1.6 m.y. long. Our goal is to provide, together with the companion study on Hole 806B (Berger et al., this volume), a more detailed and reliable Quaternary oxygen isotope stratigraphy for the western equatorial Pacific than has been available. The record of Site 805 should add information on the effects of dissolution on the δ 18 θ stratigraphy. The 1 Berger, W.H., Kroenke, L.W., Mayer, L.A., et al., 1993. Proc. ODP, Sci. Results, 130: College Station, TX (Ocean Drilling Program). Geological Research Division, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093, U.S.A. 3 Fachbereich Geowissenschaften, Universitat Bremen, Postfach 330440, D 2800 Bremen 33, Federal Republic of Germany. site is situated just above the regional lysocline (at about 3300 m; Berger et al., 1982), but close enough to experience the effects of dissolution, especially whenever the lysocline rises to a shallow level (Hebbeln et al., 1990; Groetsch et al., 1991; Le and Shackleton, 1992). The sedimentation rate (about 17 m/m.y.) is adequate for the resolu tion here attempted. MATERIALS AND METHODS Cores Studied The three uppermost cores at Hole 805C were taken from 0 to 7.8, 7.8 to 17.3, and 17.3 to 26.8 mbsf. Their lengths are 7.80, 9.88, and 9.40 m (recovery: 100%, 104%, and 99%, respectively). The cores contain light gray to white (downward in the section) foraminifer nannofossil ooze and nannofossil ooze with foraminifers. The sedi ment is moderately bioturbated. Mottled color banding is common and appears to be cyclic in places. Minor drilling disturbances are seen at the very top of Cores 130 805C 2H and 3H. Sample Preparation Cores 130 805C 1H through 3H were sampled at 10 cm inter vals, from near the surface to 26.8 mbsf. Approximately 5 g of wet bulk sediment were freeze dried for each sample, weighed, and wet sieved at 63 µm. The material was exposed to ultrasound twice for about 10 s during the process. The sand fraction (>63 µm) was dried in an oven at 50°C for 40 hr, and then weighed again to determine the percent sand fraction. For each sample, 25 tests of the planktonic foraminifer taxa Globigerinoides sacculifer and Pulleniatina were picked in the 355 425 µm fraction and crushed with a glass pestle. In some cases, fewer than 25 tests were available. The number of specimens (25) and the rather narrow size fraction were chosen to minimize the influence of vital effects on the isotopic ratios (Berger et al., 1978). For G. saccu lifer, only tests that were intact were selected. An effort was made to avoid G. fistulosus where present; the immature members of the species are difficult to distinguish from mature G. "trilobus" (the 363
Transcript
Berger, W.H., Kroenke, L.W., Mayer, L.A., et al., 1993Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 130
21. QUATERNARY OXYGEN ISOTOPE RECORD OF PELAGIC FORAMINIFERS:SITE 805, ONTONG JAVA PLATEAU1
W.H. Berger,2 T. Bickert,3 H. Schmidt,3 G. Wefer,3 and M. Yasuda2
ABSTRACT
The oxygen isotope records of G. sacculifer and Pulleniatina in the uppermost three cores at Ocean Drilling ProgramHole 805C span the last 1.6 m.y., an estimate based on Fourier stratigraphy. The last 700,000 yr are dominated by both eccentricity-and obliquity-related orbital fluctuations. The range of variation of δ 1 8 θ values is about 1.5‰, of which ca. 75% may be assignedto global ice-volume effect. The remainder of the range is shared by the effects of surface temperature variation, thermoclinedepth change (in the case of Pulleniatina, especially), and differential dissolution.
Before 1 Ma, obliquity-related fluctuations dominate. The transition between obliquity- and eccentricity-dominated timeoccurs between ca. 1 and 0.7 Ma. It is marked by irregularities in phase relationships, the source of which is not clear. The ageof the Brunhes/Matuyama boundary is determined as 794,000 yr by obliquity counting. However, an age of 830,000 yr also iscompatible with the counts of both eccentricity and obliquity cycles. In the first case, Stage 19 (which contains the boundary) iscoincident with the crest of the 19th obliquity cycle, setting the first crest downcore equal to zero, and counting backward (ol9).In the second, Stage 19 coincides with o20.
No evidence was found for fluctuations related to precession (23 and 19 k.y.) rising above the noise level, using plain Fourierexpansion on the age model of the entire series.
Detailed stratigraphic comparison with the Quaternary record of Hole 806B allows the recognition of major dissolution events(which increase the difference in δ 1 8 θ values of G. sacculifer at the two sites). These occur at Stages 11-13, 16-17, and near1.5 Ma (below o33).
INTRODUCTION
The Ontong Java Plateau straddles the equator and bears a thickcover of calcareous sediments. Already these sediments have contrib-uted much information on the stable isotope stratigraphy of theQuaternary (Shackleton and Opdyke, 1973,1976; Schiffelbein, 1984;Hebbeln et al., 1990; Wu and Berger, 1991), as well as to that of theNeogene in general (Woodruff et al., 1981; Whitman and Berger,1992). Extensive work on box cores (Johnson et al., 1977; Berger etal., 1978, 1987) provides detailed background information on pres-ent-day conditions of sedimentation, back to the last glacial.
During Leg 130 of the Ocean Drilling Program (ODP), advancedhydraulic piston coring (APC) was used to retrieve a number ofundisturbed sequences within the Neogene (Fig. 1). We have reportedon the stable isotope stratigraphy of the uppermost five cores ofHole 806B elsewhere in this volume (Berger et al., this volume;Bickert et al., this volume; Schmidt et al., this volume). For compari-son with this record, we selected the first three cores from Hole 805C(1°13.69'N, 160°31.77'E; water depth, 3188 m) for detailed study ofthe Quaternary. These cores (130-805C-1H through -3H) comprisethe upper 26.8 m of sediment, which consists of calcareous ooze withforaminifers and nannofossils. Preservation is quite good on thewhole. The period represented by the sequence sampled is about1.6 m.y. long.
Our goal is to provide, together with the companion study onHole 806B (Berger et al., this volume), a more detailed and reliableQuaternary oxygen isotope stratigraphy for the western equatorialPacific than has been available. The record of Site 805 should addinformation on the effects of dissolution on the δ 1 8 θ stratigraphy. The
1 Berger, W.H., Kroenke, L.W., Mayer, L.A., et al., 1993. Proc. ODP, Sci. Results,130: College Station, TX (Ocean Drilling Program).
Geological Research Division, Scripps Institution of Oceanography, University ofCalifornia, San Diego, La Jolla, CA 92093, U.S.A.
3 Fachbereich Geowissenschaften, Universitat Bremen, Postfach 330440, D-2800Bremen 33, Federal Republic of Germany.
site is situated just above the regional lysocline (at about 3300 m;Berger et al., 1982), but close enough to experience the effects ofdissolution, especially whenever the lysocline rises to a shallow level(Hebbeln et al., 1990; Groetsch et al., 1991; Le and Shackleton, 1992).The sedimentation rate (about 17 m/m.y.) is adequate for the resolu-tion here attempted.
MATERIALS AND METHODS
Cores Studied
The three uppermost cores at Hole 805C were taken from 0 to 7.8,7.8 to 17.3, and 17.3 to 26.8 mbsf. Their lengths are 7.80, 9.88, and9.40 m (recovery: 100%, 104%, and 99%, respectively). The corescontain light gray to white (downward in the section) foraminifernannofossil ooze and nannofossil ooze with foraminifers. The sedi-ment is moderately bioturbated. Mottled color banding is commonand appears to be cyclic in places. Minor drilling disturbances areseen at the very top of Cores 130-805C-2H and -3H.
Sample Preparation
Cores 130-805C-1H through -3H were sampled at 10-cm inter-vals, from near the surface to 26.8 mbsf. Approximately 5 g of wetbulk sediment were freeze-dried for each sample, weighed, andwet-sieved at 63 µm. The material was exposed to ultrasound twicefor about 10 s during the process. The sand fraction (>63 µm) wasdried in an oven at 50°C for 40 hr, and then weighed again todetermine the percent sand fraction.
For each sample, 25 tests of the planktonic foraminifer taxaGlobigerinoides sacculifer and Pulleniatina were picked in the 355-425 µm fraction and crushed with a glass pestle. In some cases, fewerthan 25 tests were available. The number of specimens (25) and therather narrow size fraction were chosen to minimize the influence ofvital effects on the isotopic ratios (Berger et al., 1978). For G. saccu-lifer, only tests that were intact were selected. An effort was made toavoid G. fistulosus where present; the immature members of thespecies are difficult to distinguish from mature G. "trilobus" (the
363
W.H. BERGER ET AL.
t Nauru'
155° 160°Figure 1. Track of Leg 130 on Ontong Java Plateau and position of sites (from Berger et al., 1991).
165°E
364
QUATERNARY OXYGEN ISOTOPE RECORD, SITE 805
non-"sac" phenotype of G. sacculifer). The sample size for isotopicmeasurements was 60-80 µg; the carbonate was reacted with phos-phoric acid at 75°C. Isotopic ratios were determined using a FinniganMAT 251 Micromass spectrometer with a Finnigan Automated Car-bonate Device, at the Geoscience Department of the University ofBremen. Precision was regularly checked by running standards(Solnhofen Limestone). Over a 1-yr period (1990), standard devia-tions were <0.07‰ for δ 1 8 θ and 0.5%o for δ13C. Conversion to thePeeDee Belemnite (PDB) scale was performed by means of theNBS-18, NBS-19, and NBS-20 standards.
Interpretation of Oxygen Isotopes
The methodology of interpreting the oxygen isotopes in pelagicforaminifers is summarized in Berger and Gardner (1975) and Weferand Berger (1991). The basis for interpretation is the well-knownpaleotemperature equation of Epstein and co-workers (1953):
(1)
where t is temperature, δ s is the δ 1 8 θ value of the solid carbonate, andA is the δ 1 8 θ value of the water in which the carbonate precipitated,measured in the PDB system (Epstein and Mayeda, 1953; Epstein etal., 1953). The δ value is given as a proportional deviation of isotopicratios from a standard:
δ(m) = (Rm-Rs)/Rs,
where R is the ratio of the heavier to the lighter isotope, Rm denotesmeasured, and Rs denotes standard. The value for δ(m) is given inpermil, that is, 1000 times the actual value of the deviation.
Term A in Equation 1 is a function mainly of the amount of waterlocked up in continental ice and the isotopic composition of that ice.Also, it contains information about fractionation processes involvingevaporation and precipitation at the sea surface. These processessimultaneously influence salinity, so that an overall correlation existsbetween salinity and isotopic composition of seawater. The effect canbe estimated and absorbed into the coefficient of slope in Equation 1,so that A then stands for the glacial effect alone (Berger and Mayer,1987; Whitman and Berger, 1992). For this purpose, the slope coef-ficient in Equation 1 is set to 5.0, that is, a 1°C change in temperatureproduces a 0.2‰ change in the oxygen isotope ratio of foraminiferaltests. In the region studied here, we expect that the temperature rangebetween glacials and interglacials is kept small by powerful feedbackmechanisms involving evaporation and cloud shading (Ramanathanet al., 1989). From the pattern displayed in the CLIMAP reconstruc-tions (CLIMAP Project Members, 1976), we assume that this rangeis about the same as the seasonal range, which is near 1 °C. Thus, 0.2‰of the 1.3‰ -1.5‰ range observed for G. sacculifer may be ascribedto temperature, leaving 1.1‰-1.3‰ for other factors. An effect ofdifferential dissolution on the range is likely, on the order of 0. l‰ to0.2‰. This leaves between l.l‰ and 1.2‰ for the range of the iceeffect. Confidence in this assumption, of course, weakens as one pro-ceeds backward into the record.
Depth Assignments
Depth assignments of samples are based on driller's depth and thedistance of the sampling level from the core top within the recoveredcore (see ODP depths [given in mbsf], as listed in the individual sitechapters of the Initial Reports volume). The ODP depths are adjustedby multiplying values for each core by a factor of <l .0 to account forexcess recovery (>100%) and to make room for the gap postulated atthe end of each core (estimated at 30 cm each; see below). Thisprocedure results in an "adjusted" depth below seafloor (ambsf). For
interpolation (5-cm intervals), records are first "filled in" to 2.5-cmintervals in the ODP scale, using estimates both from linear interpo-lation of adjacent measurements and from extrapolation of slopesfrom adjacent points. The final assignments are by straight lineinterpolation, between adjacent points of the artificially "filled-in"record ("resampling").
Age Assignments
Age assignments are based on Fourier analysis of the record ofeach core. Based on a preliminary estimate for the sedimentation rateof 17 m/m.y. (see Shipboard Scientific Party, 1991, "Site 805" chap-ter), the terms in the Fourier expansion that contain fluctuations in thevicinity of 41 k.y. (the period of obliquity oscillations in the EarüYsaxis) can be identified. These terms are then used to synthesize theapparent "obliquity response" in the sedimentary record. The result-ing fluctuations are assigned a distance of 41 k.y. from peak to peak,which yields an estimate for the instantaneous sedimentation rate(ISR) within each meter of the section. A smooth curve is fit throughthese rates, using low-order Fourier terms. A summation of timeintervals for successive depth intervals then yields an age for eachdepth in the core. No assumptions are made in the analysis beyondthe one that the 41-k.y. obliquity cycle should be represented in theisotopic record. This assumption is abundantly supported by previouswork (Hays et al., 1976; Pisias, 1976; Imbrie et al., 1984). Stabilityof the orbital signals seems assured for this interval (Berger, 1984).For details on the method, see the companion paper on Site 806(Berger et al., this volume).
RESULTS AND DISCUSSION
Raw Data and Overview
Results of isotopic analyses are given in Table 1 and plotted inFigures 2A and 2B. The oxygen isotopes of G. sacculifer andPulleniatina show the types of fluctuations, which are familiar fromprevious work on the plateau (Shackleton and Opdyke, 1973,1976),with large amplitudes in Core 130-805C-1H, intermediate ones inCore 130-805C-2H, and rather small ones in Core 13O-8O5C-3H.Also, typical wave lengths for the variations are seen to vary fromabout 1.5 m in Core 130-805C-1H to <l m in Core 13O-8O5C-3H,with a mixture of long and short waves in Core 130-805C-2H. Thistrend fits with previous findings that the late Quaternary record isdominated by a 100-k.y. cycle (eccentricity of Earth's orbit), and theearly Quaternary by a much shorter cycle, corresponding to obliquity(Shackleton and Opdyke, 1976,1977; Pisias and Moore, 1981; Rud-diman et al., 1986). The range in Core 13O-8O5C-1H (late Quater-nary) is near 1.5‰ (4 times standard deviation). If we assume 1° oftemperature variation, this yields a range of 1.3%ofor non-tempera-ture factors. Dissolution effects may be responsible, in part, forincreasing this range over the equivalent one in Hole 806B duringthe same time interval (1.3‰for 4 times standard deviation; Bergeret al., this volume).
Typically, values of δ 1 8 θ are slightly more positive in Core 130-8O5C-1H than in Core 130-805C-3H; that is, the average ice mass isgreater in the late Quaternary, or there is a cooling trend, or both. Thistrend is stronger in Pulleniatina than in G. sacculifer, suggesting thatthe thermocline tends to rise from early to late Quaternary (or thatPulleniatina changes its habitat to a deeper level, on average). Effectsof differential dissolution may also enter into the equation: for in-creased dissolution, δ 1 8 θ values become more positive (Wu and Ber-ger, 1989). Ranges of δ 1 8 θ values are substantially less in Core 130-805C-3H than in Cores 130-805C-2H and - 1H. It appears that the long-wave effect (eccentricity) that arises in mid-Quaternary time may beconsidered an add-on to the short-wave effect (obliquity), so that theamplitude is correspondingly increased.
365
W.H. BERGER ET AL.
Table 1. Oxygen isotopes in G. sacculifer and Pulleniatina, Cores 130-805C-1H, -2H, and -3H.
Notes: Depths are uncorrected official ODP values(Kroenke, Berger, Janecek, et al., 1991). Oxygen isotopesare given as negative δ18O‰ (deviation from PDBstandard).
DEPTH MODEL AND DEPTH SERIES
General
The depths given for the raw data (Figs. 2A and 2B) are the ODPdepths, which are based on counting core barrels, by the driller, andon measuring depth-in-core after retrieval (mbsf). It is to be expectedthat there might be gaps between cores as well as an overlap betweenadjacent cores, because of core expansion on deck. To obtain a bestestimate for depth below seafloor for each sample, we make thefollowing assumptions:
1. The top of the sediment in the first core is taken as 0.0 mbsf.2. The top of all other cores is the driller's depth (ODP depth).3. Cores must not overlap; thus, the minimum gap between the last
sample of one core and the first sample in the next below is the distanceof the two samples that appears when the bottom of the first core is laidflush against the top of the second. This is the sampling gap.
4. A coring gap is expected; it must be deduced by (a) matchingproperties of multiple cores (with breaks at different horizons), or(b) matching of properties with an idealized profile, or some otherprofile based on expectations from previous work.
We have used two methods to find any coring gaps: (1) thematching of the δ 1 8 θ record of Hole 805 C with those of CoresV28-238 and V28-239, from the same area (published by Shackletonand Opdyke, 1973,1976); and (2) a visual comparison of gamma-rayattenuation porosity evaluator (GRAPE) data for Holes 805 A, 805B,and 805C, as taken on board (kindly provided by T.R. Janecek). Theattempt at matching patterns of GRAPE series turned out to beunconvincing, because correlations between the parallel cores are notconsistently obvious. However, GRAPE data do support our results
from matching with the Vema records: all estimates for the coringgaps (Cores 1H to 2H, 2H to 3H, and 3H to 4H) came to within a fewcentimeters of 30 cm. Hence, this is the estimate we use here.Derivation of the gap size is summarized in Table 2.
Once a gap has been identified, it needs to be filled in withestimates of isotopic values, so that evenly spaced series can beextracted from the record for analysis. Straight-line interpolationbetween the boundaries of the gap is unattractive. This becomesobvious when the points at the edges are unusually high or low. In theabsence of other information, it is more reasonable to assume thatthere will be two factors allowing extrapolation from the edges intothe gap: a tendency for continuation of an established trend onapproaching the gap, and a tendency to return toward the mean. Therecords treated here have strong cyclic elements and are autocorrela-tive at a shift of around 200 k.y. (two eccentricity cycles and fiveobliquity cycles). This autocorrelation is useful when guessing miss-ing values within coring or sampling gaps. In filling gaps, we haveused these three methods (continuity of trend, return to average, andautocorrelation) with roughly equal weight. We eschewed more so-phisticated mathematical techniques as needless complications thatprovide a false sense of security. It is well to remember that, whateverthe approach, the filling of gaps is strictly guesswork. We intend toimprove on these estimates in the future by additional sampling andsplicing from other records.
Match Between Hole 805C and Core V28-238
Hole 805C was drilled at a depth of 3188 m, and Core V28-238was taken at 3120 m, only a few miles away. Thus, one would expectan excellent match between the two records. On the whole, thisexpectation is fulfilled (Fig. 3). The match suggests, however, thatone of the two cores is disturbed within certain intervals. We thinkthat the more reliable stratigraphy of the two cores is the one fromHole 805C.
The match between the two cores (Fig. 3) shows the originalrecord of Core V28-238 on top (with a linear depth-scale transform),and the adjusted record on the bottom (V28-238 altered). The recordof Hole 805C represents the δ 1 8 θ stratigraphy for Cores 130-805C-1H and -2H, with the gap (coring plus sampling) filled as shown. TheODP depths for Core 130-805C-1H were multiplied by 0.96 to obtainroom for the gap (and to take account of excess recovery). Measureddepths within Core 130-805C-2H were multiplied by 0.93 and addedto 7.80 mbsf (the driller's depth for the core top). The Brunhes/Ma-tuyama boundary in Hole 805C (ODP depth, 12.7 mbsf; ShipboardScientific Party, 1991) then appears at 12.36 mbsf, as shown. Thesampling depths given for Core V28-238 were consequently multi-plied by a factor of 1.03 (i.e., 12.36/12.00) to adjust for the differencein depth for this boundary (12 m in V28-238).
One can see that the numbered isotope stages of Core V28-238are readily matched to the corresponding excursions in the record ofHole 805C. However, compared with Hole 8O5C, the section downto Stage 13 seems expanded (especially so for Stage 11), and thesection between Stages 15 and 19 seems greatly reduced. The mis-match around Stage 11 is addressed by Prell et al. (1986, p. 149), whonote that the section around Stage 11 in Core V28-238 was disturbedby uncoupling of the core pipes. Some disturbance is also postulatedfor the structure within Stage 5, from the same process, "when thecore was extruded." Presumably, similar disturbances may exist inother portions of this core. Visual comparison of the normalizedstratigraphy of nearby Core RC17-177 with that of Core V28-238(given in Prell et al., 1986, in fig. 9) suggests that Stages 13 and 15are correctly represented in Core V28-238 (as is necessary if we areto use this record for finding the gap in Hole 8O5C), but that Stage 17is indeed underrepresented (as suggested by the match in Fig. 3).
The record at the bottom of Figure 3 represents the δ 1 8 θ stratigra-phy of Core V28-238, adjusted for a fit to the one of Hole 805C. Theadjustment was made by expanding and compressing 1-m sections of
367
W.H. BERGER ET AL.
G. sacculifer
0 1 2 3 4 5 6 7 8 9 10Distance from core top (m)
é 3
o1 2f
Pulleniatina
130-805C-1H
-3H
0 1 2 3 4 5 6 7 8 9 10
Distance from core top (m)
Figure 2. Oxygen isotope records of G. sacculifer (A) and Pulleniatina (B), raw data. Gaps are shown where the distance between samples is >15 cm.
Table 2. Depth-related parameters (in meters) forCores 130-805C-1H to -3H.
Core
Length coredRecoveredODP depthsFirst sampleLast sampleSampling gapCoring gap (V28)Coring gap (GRP)FactorAccepted coring gapAdjusted depth
1H
7.807.80
0-7.80.107.49
0.31+0.10.250.360.960.30
0-7.5
2H
9.509.88
7.8-17.30.109.70
0.18 + 0.10.370.320.930.30
7.8-17.0
3H
9.509.40
17.3-26.80.109.19
0.21 +0.1n.d.0.300.980.30
17.3-26.5
Note: Also see note added in proofs after "Acknowledgments"section (this chapter).
the original record shown on top. The corresponding factors represent aseries of sedimentation rate ratios (SRR), which are plotted in Figure 4.The ratio starts at 1.2 at the top and drops to near 1 below the Brunhes/Matuyama boundary. Distinct excursions occur within Stages 6 and 11,in rough agreement with the disturbances identified by Prell et al. (1986).An excursion to low values within Stage 16 is present, which suggestsyet another disturbance in the record of Core V28-238.
In summary, the match with Core V28-238 allows identificationof a gap of about 30 cm between Cores 13O-8O5C-1H and -2H, andalso allows assignment of isotope stages. Two out of three discrepan-cies between the two cores are identified as probable disturbances inCore V28-238. The question of whether Stage 16 is undersampled inCore V28-238 or oversampled in Hole 805C is addressed below.
To find the gap between Cores 130-805C-2H and -3H, and tofurther document the nature of discrepancies between the records ofHole 805C and Core V28-238, we next match the stratigraphies ofHole 8O5C and Core V28-239 (as given in Shackleton and Opdyke,1976) (Fig. 5). The procedure is that described for Figure 3. The linearadjustment of the depths of Core V28-239, by correlation to theBrunhes/Matuyama boundary, produces an excellent fit with Hole805C down to Stage 17. Stage 19 is reduced or missing in CoreV28-239; this is the stage that has the Brunhes/Matuyama (B/M)boundary. Stage 21 also seems reduced in Core V28-239, whereasStage 23 is well represented. Stretching the Core V28-239 recordbelow the Brunhes/Matuyama boundary to make it fit the oxygenisotope record of Hole 8O5C (bottom series) also brings in line theMatuyama/Jaramillo boundary positions.
The relationships between the two stratigraphies are plotted inFigure 6 as instantaneous sedimentation rate ratios. Core V28-239comes from a depth of 3490 m and is well off the equator (3°15'N).
Thus, its sedimentation rate is lower than that of Hole 805C and theSRR is below 1. It starts near 0.7 at the top, decreases toward theMatuyama/Jaramillo boundary, rises to a little over 0.8 below that,and drops again below 14 m in the core. There are two majordownward excursions. The first, going down in Core V28-239, is atStage 11. This is the reverse of the situation in Core V28-238. Thus,Hole 805C represents the intermediate (and most likely true) condi-tion. The second excursion, which is quite large, is in the vicinity ofthe Brunhes/Matuyama boundary. Also, one can see that no anomalyis present at Stage 16, so that the records of Core V28-239 and Hole805C agree that this section is underrepresented in Core V28-238.
In summary, when comparing the SRR anomalies of Core V28-238/Hole 805C and Core V28-239/Hole 805C, one can see that theirpatterns are not similar, which would be the case if Hole 8O5C hadthe odd record. This supports our contention that the depth model forHole 805C is essentially correct, and that its stratigraphy is morereliable than that of either of the Vema cores.
AGE MODEL AND TIME SERIES
Dating by Counting Obliquity Cycles:G. sacculifer Record
The assignment of ages, as mentioned, is based on the counting ofδ I 8 0 cycles of approximately 40-k.y. length (i.e., typically 60-80 cmwavelength in the cored sequence). Equally spaced 518O values wereobtained by first filling in between adjacent samples using a weightedmean of three estimates: the mid-point of the straight-line interpola-tion, and two values derived from extrapolating the trends from thetwo points preceding the new position and the two following it.Giving the first estimate 4.5 times the weight of the others results ina smooth sinusoidal fit. This new curve is then resampled at 5-cmintervals by straight-line interpolation. The Fourier components of theseries were determined separately for three sections: (1) the lateQuaternary or "Milankovitch Regime," in which eccentricity cyclesdominate; (2) the middle Quaternary or "Croll Regime," in whichboth eccentricity and obliquity are important, and (3) the early Qua-ternary or "Laplace Regime," in which obliquity cycles dominateentirely. The terms are chosen for historic and mnemonic reasons(Berger and Wefer, in press).
The Milankovitch Regime is defined as the sequence from thepresent back to δ 1 8 θ Stage 15 (or, more precisely, the crest of the 15thobliquity cycle, counting the last crest as zero) (Fig. 7). The top of therecord was extended by shifting the last 200 k.y. forward into thefuture. The appendix is linearly tapered toward the mean, at 200 k.y.into the future (not shown). The assumed sedimentation rate is1.4 cm/k.y. The appendix at the end of the series was formed inanalogous fashion. Figure 7 shows the extended record, a Fourier fit
Figure 3. Comparison of δ 1 8 θ stratigraphies of Hole 805C (this study) andCore V28-238 (Shackleton and Opdyke, 1973). Scales shifted for clarity. Top,depth in Core V28-238 adjusted linearly to that in Hole 805C (factor = 1.03).Bottom, depth in Core V28-238 altered by mapping its δ 1 8 θ signal onto thatof Hole 805C, at 1-m intervals. Isotope stages are as given in Shackleton andOpdyke (1973).
derived from the main 12 harmonics (in essence, a smoothed versionof the original curve), an eccentricity-related component, and anobliquity-related component, all offset from each other for clarity.
The Fourier fit demonstrates that relatively few harmonics repre-sent most of the information contained in the record. The eccentric-ity-related component is represented by Harmonics 8 through 12 ofa record that is 1375 cm long; that is, it contains the cycles withwavelengths between 170 and 110 cm, approximately. This range isset to contain the eccentricity signal for sedimentation rates between1.1 and 1.7 cm/k.y. Recognition of the dated Stage 5e (see Shackletonand Opdyke, 1973) yields a first guess for the sedimentation ratebetween 1.3 and 1.4 cm/k.y. Eccentricity Cycle 6 ("E6") is seen tocoincide with Stage 15. Thus, the Milankovitch Regime is somewhatlonger than 600 k.y.
For detail, we turn to the obliquity record (Harmonics 20 to 30,corresponding to wavelengths of about 70^45 cm, for assumed sedi-mentation rates between 1.7 and 1.1 cm/k.y.). We see that the sharptransition at the beginning of Stage 15 is due to the presence of thecrest of obliquity Cycle 15 ("ol5"). The time interval from this pointto the top of the core is 15 × 41 k.y. + 15 k.y., which is 630 k.y. Theinterval indicated by counting eccentricity cycles is similar: about625 k.y. In Hole 806B the equivalent level was found as 623 ka(Berger et al., this volume). Both determinations agree well with thatgiven in the SPECMAP model (617 ka; Imbrie et al., 1984) (althoughat greater depths our scale begins to deviate markedly fromSPECMAP). Here we take the end of the transition between Stages 16and 15 as 630 yr, and this then is the duration of the MilankovitchRegime based on the present data set.
A similar analysis is next made for the Croll Regime (Fig. 8). It isdefined as the interval between ol5 and o30; that is, it is 615 k.y. long.For this series, extensions were taken from the adjacent sections andtapered linearly toward the mean at the ends. The eccentricity-relatedcomponent weakens considerably in the early part of the Croll Re-gime. Stage 23, in the middle of the Croll section, is seen to be about1 m.y. old, as it coincides with E10. It also coincides with o23, whichyields an age of 8 × 41 + 630, that is, 958 ka. The discrepancy is 42 ka,or one obliquity cycle. If we count the blip showing below ol6 in thebottom curve of Figure 8, this discrepancy disappears. If we do notcount the subdued eccentricity cycle below E7, another discrepancyarises. Thus, we could reasonably adjust the count by adding oneobliquity cycle.
There is one important problem with this adjustment, however.Recall that the Brunhes/Matuyama boundary intersects the main peak
2 4 6 8 10 12 14 16 18Depth in Core V28-238 (m)
Figure 4. Instantaneous sedimentation rate ratios between Core V28-238 andHole 805C (d[V28-238]/d[805C]), based on the fit of δ 1 8 θ stratigraphiesshown in Figure 3 (V28-238 altered). The record of Core V28-238 (SMOWscale on right) is given for identification of positions of discrepancies. Upwardexcursions indicate expansion relative to Hole 805C; downward excursionsindicate reduction of equivalent sections.
Figure 5. Comparison of δ 1 8 θ stratigraphies of Hole 805C (this study) andCore V28-239 (Shackleton and Opdyke, 1976). Scales shifted for clarity. Top,linear depth transform (matching Brunhes/Matuyama boundary; factor = 1.7).Bottom, 1-m interval mapping (as in Fig. 3).
(the earliest one) of Stage 19, as shown in Figure 3. A date of 730 kais conventionally assigned to this boundary (e.g., Imbrie et al., 1984;Prell et al., 1986; Ruddiman et al., 1986; Raymo et al., 1990); but thishas recently been challenged by the proposition that the age shouldbe near 780 ka (Shackleton et al., 1990; Baksi et al., 1991, 1992; seealso Johnson, 1982; Izett and Obradovich, 1991). The lesser agewould put Stage 19 just above ol8 in the section. The greater agewould correlate it exactly with ol9. If we count the blip below ol6as a cycle, we would then arrive at o20 for correlation with Stage 19,assigning an age of 830 ka to the Brunhes/Matuyama boundary. Thisage is well outside the range of values previously proposed. Forpresent purposes, we shall ignore the blip and count as shown. Thisputs the Brunhes-Matuyama boundary into ol9, for an age of 4 × 41+ 630 (i.e., 794 ka), in good agreement with the preference ofShackleton et al. (1990) and with the age model for Hole 806B (Bergeret al., this volume).
Nevertheless, it is unsatisfactory that o23 should coincide withE10, because this implies that the usual phase relationships betweenthe major cycles are not valid in this vicinity. In other words,Stages 19, 21, and 23 do not fit the pattern seen in the more regular
369
W.H. BERGER ET AL.
0 2 4 6 8 10 12 14 16Depth in Core V28-239 (m)
Figure 6. Instantaneous sedimentation rate ratios between Core V28-239 andHole 805C (d[V28-239]/d[805C]), based on the fit of δ 1 8 θ stratigraphiesshown in Figure 5. The record of Core V28-239 is given for identification ofpositions of discrepancies (PDB scale on right). Upward excursions indicateexpansion relative to average ratio to equivalent sections in Hole 805C;downward excursions indicate reduction.
Milankovitch Regime. The discrepancy arises because E8 is so poorlyexpressed, which in turn is due to the fact that Stage 18 is weak andshort. The composite δ 1 8 θ record of Prell et al. (1986) suggests thatthis is generally true and not just peculiar to the record of Hole 805C.The possibility that the proposed age of the Brunhes/Matuyamaboundary, at 794 ka, is still too young by one obliquity cycle cannotbe discounted. The date for this boundary has seen many reevalu-ations for the last 20 yr toward higher values.
Fourier analysis of the Laplace Regime, defined as the intervalbetween o30 and o45, demonstrates the dominance of the obliquitycycles, and the absence, in essence, of eccentricity cycles (Fig. 9). Theavailable series is seen to bottom in o39; self-similarity was used toappend the extension, which was tapered to avoid edge effects in theFourier components. The range of harmonics used for synthesis ofeccentricity- and obliquity-related variations ("window") was 7 to 12and 19 to 30, respectively, as before.
Results of the counting exercise are summarized in Figure 10 forready reference. Counting of traditional isotope stages, in the odd-even scheme introduced by Emiliani (1955), ends with Stage 23(Shackleton and Opdyke, 1976). We have not attempted to label olderstages by comparing with other stratigraphies (Ruddiman et al., 1986;Shackleton et al., 1990; Raymo et al., 1990), because such correlationrequires careful matching of biostratigraphic information. (The nec-essary data are now available through the detailed investigations ofT. Takayama [this volume], so that these comparisons will be possiblein the near future.) Instead, we have used the regular sequence ofobliquity cycles, which allows instant conversion to an age estimate.The point at which the crest of each cycle appears is dated relative toObliquity Crest 0, which is the very first crest just below the core top("oO"). The time interval from oO to the top is found by extrapolatingthe sedimentation rate between oO and the next deeper crest ("ol")toward the top of the core.
How reliable is our method of dating? There are several ways toapproach this question: internal consistency within the same record(comparison of eccentricity- and obliquity-related signal), compari-son with results from another record within the same core, andcomparison with other cores. We have mentioned the good agreementwith the SPECMAP age model back to Stage 15. Also, we havereferred to internal consistency earlier and found it somewhat lackingin the vicinity of the Brunhes/Matuyama boundary. Apparently, in theinterval of transition to a dominance of major eccentricity cycles, thesystem does not behave in the highly regular fashion that is apparentbefore and after this period.
-2 2 4 6 8Adjusted depth (ambsf)
10
Croll
15
V
y \15
Ml
A (\Λ
7 9
AX Λ
WVw
23
AS ΛΛΛA A
Λ/v y i
Λ
10
22 23 Λ Λ
VW\/\/\
A A Λ,Λ A
Λ /λ
ΛΛΛM
G. sacculifer
AMraw
main 13
ca. 100 k.y.
Figure 7. Fourier analysis of the upper Quaternary portion of the δ 1 8 θ recordof G. sacculifer in Hole 805C, extended by shifting autocorrelated ends andtapered (see text). Milankovitch Regime: last 15 obliquity cycles. "Main 12"refers to dominant harmonics.
Figure 8. Fourier analysis of the middle Quaternary portion of the δ 1 8 θ recordof G. sacculifer in Hole 805C, extended by shifting ends and tapered (see text).Croll Regime: period between 15th and 30th obliquity cycle. "Main 13" refersto dominant harmonics.
We next compare the results so far obtained by a similar analysisfor the δ 1 8 θ record of Pulleniatina, to further check on reliability.
Dating by Counting Obliquity Cycles:Pulleniatina Record
The procedure of analysis is the same as that for G. sacculifer:Figs. 11, 12, and 13 correspond to Figs. 7, 8, and 9, respectively. Inthe Milankovitch portion (Fig. 11), the most striking aspect is againthe dominance of the eccentricity-related cycles. There is a problemin Stage 15: both the 100 and 41 k.y. windows show poor fit to thedata. Apparently, Stage 15 is too broad for a good fit; that is, thetransition from Stage 16 to Stage 15 starts too early or is otherwiseirregular. The Croll portion (Fig. 12) pretty much repeats the pictureseen in the G. sacculifer record—again, there is the mismatch be-tween E10 and o23, and the "blip" below ol6 that, if counted as acycle, would remove the discrepancy. In the Laplace section of thePulleniatina record, the eccentricity-related cycles are slightly greaterthan those seen in G. sacculifer, but they still have very little power.
The summary graph for obliquity counting in the Pulleniatinaδ 1 8 θ record (Fig. 14) is rather similar to the one for G. sacculifer(Fig. 10). Stage 23 is seen to correspond to o23, and the recordterminates just before o38, within o39.
370
QUATERNARY OXYGEN ISOTOPE RECORD, SITE 805
039
G. sacculifer
Laplace
18 20 22 24 26Adjusted depth (ambsf)
main 13
ca. 100 k.y.
ca. 41 k.y.
28 30
Figure 9. Fourier analysis of the lower Quaternary portion of the δ 1 8 θ recordof G. sacculifer in Hole 805C, extended by shifting ends and tapered (see text).Laplace Regime: period between 30th and 45th obliquity cycle. "Main 13"refers to dominant harmonics.
Figure 10. Summary of obliquity cycle counting in the δ 1 8 θ record of G. sac-culifer in Hole 805C, down to obliquity cycle 39. "o33" = crest of 33rdobliquity cycle. (The most recent crest is defined as zero.)
A detailed comparison of the results of obliquity dating for thetwo records emerges when plotting instantaneous sedimentationrates (ISR) on top of each other (Fig. 15). Over much of the sequencethe agreement is excellent. Sedimentation rates typically vary be-tween 1.3 cm/k.y. near the top of the core and 2 cm/k.y. in the middleQuaternary. Low values are shown in the vicinity of the Brunhes/Ma-tuyama boundary. The lower Quaternary shows somewhat lower ISRvalues in both records. However, a section with major discrepanciesis present in the vicinity of Stages 13 to 15. The section of greatestdisagreement coincides with the break between Cores 130-805C-1Hand -2H; this may account for much of the problem. However, thereis strong disagreement also at the beginning of Stage 15, which is wellbelow the break.
Age Assignments and Age Model
There seems to be no good reason to prefer the estimates ofinstantaneous sedimentation rate results of one taxon over that fromthe other; therefore, for the subsequent age assignment, we have takenthe average of the two independent ISR determinations.
Using this average of the smooth curves shown in Figure 15, wecan now assign an age to each position in Hole 805C. We did this at5-cm steps, and then we resampled at intervals of 4 k.y. by straight-line interpolation (see Table 3). The resulting curves represent the agemodel for Cores 130-805C-1H through -3H (Fig. 16).
The entire sequence was then analyzed once again in the timedomain (Fig. 17). The windows are set at 87-118 k.y. and 36-47 k.y.,that is, about 15% beyond either side of the expected periods. Theresults confirm in some detail what has emerged already: (1) theexcellent agreement between the records of G. sacculifer and Pulleni-atina; (2) the dominance of the eccentricity-related cycle in theinterval since ol5 (Milankovitch Regime); (3) the dominance ofobliquity-related fluctuations in the interval below o27 (lowermostCroll and Laplace regimes); and (4) the transitional interval betweeno27 and ol5 (most of the Croll Regime). The irregularity of theinterval in the vicinity of Stages 17 to 19 is reflected in the apparentphase reversal of the eccentricity-related cycles.
The fact that the eccentricity-related cycles are not entirely"clean" is also seen in the Fourier spectrum (Fig. 18). The 100-k.y.peak has a lesser but distinct adjunct to the right (higher harmonic),which produces an interference pattern. The 41-k.y. peak, naturally,is strong and distinct as a result of our procedure. The Pulleniatinarecord has a Fourier peak near 30 ka, which is absent from theG. sacculifer record. Power within the band corresponding to pre-cession (19 and 23 k.y.) is but slightly elevated over background: the
1
Λ A/ V y x / V
" 0
0
Milankovitch
kV
r~\
" \
/A/̂ \
\ Λ Λ
7 9 1J
U AM /I
y J \\ V5 10
13
Λ
r5\
1 -
Pulleniatina
A M ^T raw
A/-V/main 12
ca. 100 k.y.
15
ca. 41 k.y.
-2 2 4 6 8Adjusted depth (ambsf)
10 12
Figure 11. Fourier analysis of the upper Quaternary portion of the δ1 8O record ofPulleniatina in Hole 805C (Milankovitch Regime). Compare with Figure 7.
present analysis presents no evidence on the importance of precessionin these records.
COMPARISON OF OXYGEN ISOTOPE RECORDSWITHIN HOLE 805C
Eccentricity Domain
We have noted earlier, when discussing Figures 2 and 3, that theδ 1 8 θ records of G. sacculifer and Pulleniatina show certain differ-ences that change through time. In Core 13O-8O5C-1H, the δ 1 8 θvalues of the two taxa differ by 0.42%e on average, in Core 13O-8O5C-2H by 0.29‰, and in Core 130-805C-3H by 0.27‰. Thus, there isan increase in difference with time within the Quaternary. Also, theamplitude of the δ 1 8 θ fluctuations increases: from 0.8‰ in Core13O-8O5C-3H to ca. 1.5‰ in Cores 130-805C-2H and -1H. Couldthere be a relationship between these two trends? Differential disso-lution decreases the contrast between δ 1 8 θ values of G. sacculifer andPulleniatina (Wu and Berger, 1989). Thus, the relatively smallercontrast in Core 130-805C-2H, despite higher overall amplitudes,may in part be due to increased dissolution effects in that core.
Before such questions can be properly attacked, the actual patternsof the relationships between the δ 1 8 θ records of the two taxa must beestablished in some detail. In the following, we derive averagepatterns over one eccentricity cycle by summing the relevant variablesmodulo 100 k.y.; that is, we "stack" successive eccentricity cycles
Figure 12. Fourier analysis of the middle Quaternary portion of the δ 1 8 θ record
of Pulleniatina in Hole 805C (Croll Regime). Compare with Figure 8.
Pulleniatina
o30 035 039
18 22 24 26Adjusted depth (ambsf)
main 13
ca. 100 k.y.
ca. 41 k.y.
28 30
Figure 13. Fourier analysis of the lower Quaternary portion of the δ 1 8 θ record
of Pulleniatina in Hole 805C (Laplace Regime). Compare with Figure 9.
("autostacking"). Each "autostack" consists of the average of fiveconsecutive intervals.
The first such autostack, from 0 to 500 ka, shows a distinct cycle,as expected (Fig. 19A). The peak appears at 20 to 30 ka for bothG. sacculifer ("SOX") and Pulleniatina ("POX"). Because the twocurves do not line up exactly, the difference varies: it is greatest in theearly phase of the transition from glacial to postglacial condition("deglaciation," 70-30 ka) and lowest during the transition from theinterglacial to the glacial condition ("reglaciation," 100-70 ka). Thesense of the difference variation is as expected, if dissolution plays arole. During deglaciation preservation increases, and hence the origi-nal isotopic difference is preserved. During reglaciation a dissolutionevent occurs, and hence the more susceptible thin-shelled specimensin G. sacculifer are dissolved. Because this vulnerable portion of thespecies assemblage carries a "light" δ 1 8 θ signal, the result of selectivedestruction is an approach of the average G. sacculifer values towardthe average values of Pulleniatina.
This tentative explanation of the difference pattern, however, doesnot agree with the sedimentation rate pattern. The ISR (top curve inFig. 19) is at a minimum during deglaciation (when preservation isthought to be best), and is high during reglaciation (when dissolutionis relatively strong). Thus, other factors enter also, presumably vari-ations in productivity and thermocline motions. We cannot, withoutdetailed analysis, offer ready explanations for the patterns seen.
Proceeding with the pattern inventory, we note that the five-cycleautostack for 500-1000 ka (Fig. 19B) is far less regular than the onein Figure 19A. The reason is the phase problem that arises betweenobliquity and eccentricity, as discussed above. The main peak nowappears at 60-80 ka rather than at 20-30 ka (where a remnant peakis present, however). No clear pattern emerges in the difference (DIF)or ISR curves. The lowermost autostack (Fig. 19C) shows littleevidence for the presence of an eccentricity cycle. However, the ISRrecord, interestingly, does seem to contain such a cycle. Thus, somepart of the system (one that is involved in determining accumulationrates) presumably was responding to eccentricity in the Earth's orbit,but this did not influence the δ 1 8 θ record. The diatom record (Langeand Berger, this volume) supports this interpretation.
Obliquity Domain
The autostack procedure is next applied to the 41-k.y. cycle; thatis, we form the average for six intervals of 41-k.y. length each, settingthe beginning of each interval equal to zero. Six such stacks arepresent (Figs. 20A-20F). Four out of the six (Figs. 20Aand 20D-20F)show good cycles; two (Figs. 20B and 20C) do not. The poorshowings are for the interval between 250 and 750 ka, which includesintervals where the obliquity cycles are poorly expressed (Fig. 17).
Figure 14. Summary of obliquity cycle counting in the δ 1 8 θ record of Pulleni-
atina in Hole 805C, down to obliquity cycle 38. Compare with Figure 10.
The peak for the first six cycles appears at 10-15 ka (Fig. 20A);the difference in δ 1 8 θ values is great during deglaciation (as in theeccentricity stack) and at a minimum during glacial conditions (ratherthan during reglaciation as in the eccentricity stack). The ISR is flat,as it should be, as its variations are not resolved for the obliquity cycle.It is between 1.3 and 1.4 cm/k.y. for the last 250 ka.
The cycles between 250 and 500 ka do not stack well (Fig. 20B).The trough for these cycles apparently is near 20 ka. The DIF is at amaximum during deglaciation (20-10 ka), but it is high earlier, too.The reglaciation minimum appears at roughly the expected position(40-30 ka). The ISR is near 1.6 cm/k.y. The next six cycles (Fig. 20C)are not well expressed. The DIF is at maximum during reglaciationand has no distinct minimum. Thus, it does not follow the expectedpattern of high DIF during deglaciation and low DIF during reglacia-tion. The ISR is near 1.8 cm/k.y.
The next deeper six cycles (750-1000 ka; Fig. 20C) stack nicely.The interglacial peak appears near 35 ka. Thus, the DIF maximumbetween 5 and 10 ka corresponds to deglaciation (i.e., it occurs at11-16 ka before the peak). POX clearly lags SOX by about 5 k.y. Theminimum DIF is at 32 ka, just after the peak, but it is not very wellexpressed: the only strong excursion from the background is themaximum, lasting about 15% of the cycle. The next six cycles(1000-1250 ka; Fig. 20E) again stack very well. The DIF does notchange much, and the ISR remains near 1.8 cm/k.y. The same patternemerges for the next deeper six cycles (Fig. 20F), except that ISR fallsback to 1.6 cm/k.y.
372
QUATERNARY OXYGEN ISOTOPE RECORD, SITE 805
10 15 20Adjusted depth (ambsf)
25 30
Figure 15. Instantaneous sedimentation rates (ISR) resulting from countingobliquity-related cycles in the δ 1 8 θ records of G. sacculifer and Pulleniatina.The curves shown represent low-order Fourier fits to the raw ISRs taken fromthe data in Figures 10 and 14. G. sacculifer δ 1 8 θ record shown for orientation(scale on right). Isotope stages as in Figure 3; "o33" = crest of 33rd obliquitycycle (most recent crest is set at zero). Core breaks shown by triangles.
COMPARISON OF OXYGEN ISOTOPE RECORDSHOLES 805C VS. 806B
Overview
The significance of differences in the δ 1 8 θ records of G. sacculiferand Pulleniatina in Hole 805C can only emerge when they arecompared with the corresponding records in Hole 806B, which is ata shallower depth and in which sediments have experienced muchless dissolution. A direct comparison of the oxygen isotope records,noting the differences in the values for the dissolution-sensitivespecies G. sacculifer should yield useful information on the stratigra-phy of dissolution. To this end, of course, correlation between the twosites must be on a reasonably fine time scale; otherwise, only the mostgeneral results will emerge.
Correlation of the two Quaternary signals is quite straightforwardin the upper part, where the mixture of eccentricity- and obliquity-re-lated fluctuations yields unique stratigraphic patterns (Fig. 21). How-ever, in the lower half of the Quaternary, where obliquity dominatesentirely, the very uniformity and lack of differentiating characterreduces the chances for correct correlation between peaks, unlesscomplete and undisturbed sequences can be assumed. (It might besurmised that detailed biostratigraphic correlation will eventuallyprovide the exact match between cycles. This presupposes that differ-ential dissolution in Site 805 does not affect the datum levels used.)In Figure 21, one can see that the match back to Stage 17 poses noproblems; a small gap must be assumed between Cores 130-806B- 1Hand -2H within Stage 9 (as shown). Dividing the depths of Core130-806B-1H by 1.45, and those of Core 130-806B-2H by 1.37 thenprovides the match seen. It becomes evident that Stage 19 is missingin Hole 806B, so one has to assume that quite a substantial gap existsbetween Cores 130-806B-2H and -3H. To match the record below thisgap to that of Hole 805C, the depths in Core 130-806B-3H are dividedby 1.30. One obliquity cycle seems to be missing between Cores 130-806B-3H and -4H. After adjustment, the depths for Core 130-806B-4H are divided by 1.40 to get the match shown.
Using the gap estimates for Hole 806B that are derived in thefashion illustrated, by matching the G. sacculifer δ 1 8 θ records, wecan now construct a detailed age model for the Quaternary δ 1 8 θ recordin Hole 806B by counting obliquity-related cycles, as illustratedabove for Hole 805C. After this is done (see Berger et al., this volume),the two G. sacculifer records can be compared in some detail(Fig. 22A). The difference (calculated for periods >86 k.y. only) isshown to be substantial in places (typically as much as 0.3‰ -0.4‰).
Also, the difference record is seen to contain cycles of a period closeto eccentricity. An unknown (and potentially important) portion of thepower in these cycles may be the result of small mismatches in thecorrelation, however. Wherever a slight mismatch in phase occurs,one transition (e.g., deglaciation) will show a diminished differencebetween the records, whereas the opposite one (e.g., reglaciation) willshow an increased one. Mismatch of peaks (e.g., in Stage 15) also canreverse the sign of the difference. No physical explanation is evidentfor such a reversal of sign.
To illustrate this point regarding the importance of small misalign-ments and resulting phase shifts, we compare the two G. sacculiferrecords in the eccentricity domain (Fig. 22B). The window used forsynthesis is as before (87-118 k.y.). The overall similarity of the tworecords in this band is quite striking. The Site 805 record is distinctlyoffset from the Site 806 record throughout, as might be expected fromdifferential dissolution. The phase shifts between the two recordsprovide the fluctuations seen: the mean of the difference is steady (bydefinition, as any long-term variations are filtered out). Note thechange in phase near 300 ka. It may be doubted that the correlationbetween the two records is so exact that the sense of the phase shiftscan be taken at face value. If, however, there is some reality to thephase shifts, then changes in this shift would reflect profound changesin the workings of the system: there is a great difference betweenhaving dissolution increased during deglaciation and increasing itduring reglaciation.
To what extent are the differences in δ 1 8 θ between the two records(Fig. 22A) reflected in the δ 1 8 θ record of Hole 805C alone? Onewould expect some kind of correspondence between the SOX-SOXdifference between sites and the POX-SOX difference in the deepersite, which has experienced the dissolution providing for the separa-tion of G. sacculifer values in the first place. This possibility isconsidered in Figure 23, where the difference in SOX between sitesis compared with the difference (SOX-POX) within the deeper site.The two curves correlate quite nicely in the early Quaternary, at leastin the general trends, with periods greater than eccentricity. Wheneverthe difference between sites increases (because of dissolution effects),the difference between SOX and POX in Hole 805C has a tendencyto approach zero (as expected). However, in the late Quaternary, theSOX-POX difference becomes more pronounced, despite a ratherhigh level of site-to-site difference. Thus, a primary increase in thedifference between the two taxa (a rise of the thermocline) must beassumed here.
The SOX difference signal suggests the existence of an interval ofstrong carbonate dissolution in the lower half of the Brunhes Chron. Theintervals are marked MBDI (for mid-Brunhes dissolution interval) andEBDI (for early Brunhes dissolution interval). Also, there is an indicationof a dissolution interval in the early Matuyama (labeled EMDI).
If these dissolution intervals are real, the sedimentation rate withinHole 805C should be reduced here beyond the general reduction,which is tied to the greater depth of this site. The stratigraphy ofsedimentation rate ratios (Fig. 24) does not show the expected re-sponse (a strong positive excursion during the dissolution intervals),except for the early Matuyama Event. In fact, although correlationsin the expected sense exist over much of the Quaternary, there is ahint of anticorrelation in the middle and lower Brunhes. That is,differences in sedimentation rates between the two sites decrease atthe very time when the effects of differential dissolution increase. Thisseems totally incompatible with a straightforward effect of carbonatedissolution on sedimentation rate. One explanation would be thatduring a dissolution pulse the shallower site is proportionally moreaffected than the deeper one, for example, by increased winnowingor from productivity-related effects.
SUMMARY AND CONCLUSIONS
The western equatorial Pacific is of special interest in the contextof Quaternary climate because it is a region of maximum open-ocean
373
W.H. BERGER ET AL.
Table 3. Age model for Quaternary oxygen isotope record of Hole 805C, G. sacculifer and Pulleniatina.
Notes: Derived from Table 1 by adjusting for expansion in core and coring gaps and fromage assignments based on counting obliquity-related cycles.
temperatures with very little seasonal variation. The mechanismsresponsible for keeping surface-water temperatures at rather uniformhigh temperatures throughout the year may likewise be expected tominimize variation through glacial-interglacial cycles, so that most ofthe stable isotope record in the area should reflect phenomena ofglobal significance rather than regional temperature variations. Onthe other hand, fluctuations in thermocline depth and upwelling areexpected for this area, which is part of the equatorial upwelling systemin the Pacific, which in itself constitutes a major element in globalclimate dynamics. The oxygen isotope record at Site 805, then, isdominated by a global signal of Quaternary ice-volume variation,modified by surface-water temperature fluctuations and by changesin regional thermocline depth. Also, this site is deep enough (3188 m)to be affected by carbonate dissolution, which in turn influences theoxygen isotope record.
We estimate the relative importance of the four signals (global icevolume, regional temperature, regional thermocline, deep-water satu-ration) in the present data as follows, in terms of proportion of rangeof δ 1 8 θ values controlled within the late Quaternary:
030
-0.50 200 400 600 800 1000 1200 1400 1600
Age (ka)
Figure 16. Age model for the δ 1 8 θ records of G. sacculifer and Pulleniatina
in Hole 805C, Cores 130-805C-1H through -3H. Isotope stages as in Figure 3;
"o30" = crest of 30th obliquity cycle (most recent crest is oO).
Thermocline changeDissolution change
Total
O.l‰O.l‰
1.5‰
7%7%
100%
Global ice volume l.l‰Temperature change 0.2%o
73%13%
The ice-volume effect may be slightly greater, with other factorscompensating through opposite sign. Stratigraphic separation of thesevarious effects is only possible if enough proxy variables are avail-able. It cannot be done with only the three variables considered here( δ 1 8 θ of G. sacculifer, δ 1 8 θ of Pulleniatina, and instantaneous sedi-mentation rate). Fluctuations in productivity and dissolution must beexplored using additional proxies. Time constraints prevented us fromdoing this in the present study.
The difficulties arising are illustrated by the fact that differencesin δ 1 8 θ values between G. sacculifer and Pulleniatina are a result ofchanges in mixed layer thickness as well as differential dissolution,with Pulleniatina having the deeper habitat and also being moreresistant to dissolution than G. sacculifer.
Comparisons with Hole 806B are especially useful when studyingdissolution effects on the oxygen isotope record because of theshallow position of this site (2520 m). The effect is reflected in adifference in δ 1 8 θ values for G. sacculifer, due to the fact that 18O-richshells become concentrated in Site 805 through differential removalof isotopically light specimens. Surprisingly, the variations of thisdifference between the two sites is not closely correlated with sedi-mentation rate ratios. Thus, factors other than dissolution of carbonatedominate accumulation rates in Site 805.
The main stratigraphic trends that emerge are as follows:
1. The last third of the Quaternary is characterized by strong,regular, climatic fluctuations dominated by 100- and 41-k.y. periods.
2. The early part of the Quaternary oxygen isotope record hasessentially no information in the 100-k.y. band, but it is entirelydominated by obliquity-related fluctuations.
3. The middle of the period is transitional.
There is a confusing zone between Stages 17 and 21, with an oddphase reversal in the eccentricity-related signal (E7 to E8), thatproduces a mismatch in the progression of counts based on eccentric-ity and obliquity. It is not clear whether this irregularity is a result ofproblems in recovery, in the record itself, or in the analysis attemptedhere. The problem prevents us from assigning a unique age to theBrunhes/Matuyama boundary, which is either near 790 ka or near 830ka, by our counts. We have used the more conservative lower age forour age model (794 ka, to be exact), in agreement with Shackleton etal. (1990).
Isotopic Stages 11-13 and 16-17 are characterized by large differ-ences in δ 1 8 θ of G. sacculifer at Sites 805 and 806, indicating strongdissolution. Also, differences are large below o33, near 1500 ka.
375
W.H. BERGER ET AL.
o30
1.5
1.036..47
200 400 600 800 1000 1200 1400 1600Age (ka)
Figure 17. Fourier analysis of δ 1 8 θ records of G. sacculifer and Pulleniatinain the time domain, showing eccentricity-related (87-118 k.y.) and obliquity-related (36-47 k.y.) fluctuations. In each pair, the upper curve is fromG. sacculifer, the lower from Pulleniatina. The record of G. sacculifer inHole 805C is given for orientation (arbitrary scale). "E6" = crest of sixtheccentricity cycle (most recent crest is set at zero). Crests of 15th and 30thobliquity cycles are marked.
0.12
0.10-
0.08
0.06-
0.04
0.02
n nn
FTot
A I
i
üai [41
\I ̂Ir
ka]
A
G. sacculifer
Pulleniatina
23 ka 19 ka
20 40 60 80 100Harmonic (fraction of 2012 k.y.)
120 140
Figure 18. Fourier spectrum (somewhat smoothed) of the δ 1 8 θ records ofG. sacculifer and Pulleniatina, based on analysis of the age models of theserecords, extended to 2012 k.y. (by adding mean values on both ends). Thepositions of the orbital periods (100, 41, 23, and 19 k.y.) are shown fororientation. The orbital period (41 k.y.) is the basis for the age model, so thata strong peak at 41 k.y. is expected from the dating method.
ACKNOWLEDGMENTS
We thank the members of the shipboard party of Leg 130 forassistance in sampling and other help rendered during the cruise.Dr. M. Segl, Bremen, supervised and facilitated the generation of theisotope data. Tom Janecek provided the GRAPE data used for check-ing on core breaks. We are indebted to Eystein Jansen and RichardCorfield for reading the draft manuscript and making helpful sugges-tions. Financial support was provided by the U.S. National ScienceFoundation and by the Deutsche Forschungsgemeinschaft.
Note added in proof: The nannofossil data of Takayama (thisvolume) suggest a 2-m coring gap between Cores 130-805C-2H and-3H. If real, such a gap would affect some of the conclusions regardingthat part of the record that is older than 1 m.y.
REFERENCES*
Baksi, A., Houghton, B., McWilliams, M., Tanaka, H., and Turner, G., 1991.What is the age of the Brunhes-Matuyama polarity transition? Geol. Soc.Am. Abstr. Progr., GP32B:135.
Baksi, A.K., Hsu, V, McWilliams, M.O., and Farrar, E., 1992. ^Ar/^Ar datingof the Brunhes-Matuyama geomagnetic field reversal. Science, 256:356-357.
Berger, A.L., 1984. Accuracy and frequency stability of the EaruYs orbitalelements during the Quaternary. In Berger, A., Imbrie, J., Hays, J., Kukla, G.,and Saltzman, B. (Eds.), Milankovitch and Climate (Pt. 1): Understandingthe Response to Astronomical Forcing: Dordrecht (D. Reidel), 3-39.
Berger, W.H., Bonneau, M.C., and Parker, F.L., 1982. Foraminifera on thedeep-sea floor: lysocline and dissolution rate. Oceanol. Ada, 5:249-258.
Berger, W.H., and Gardner, J.V., 1975. On the determination of Pleistocenetemperatures from planktonic foraminifera. J. Foraminiferal Res.,5:102-113.
Berger, W.H., Killingley, J.S., and Vincent, E., 1978. Stable isotopes indeep-sea carbonates: Box Core ERDC92, west equatorial Pacific.Oceanol. Acta, 1:203-216.
, 1987. Time scale of Wisconsin/Holocene transition: oxygen isotoperecord in the western equatorial Pacific. Quat. Res., 28:295-306.
Berger, W.H., Kroenke, L.W., Mayer, L.A., and Shipboard Scientific Party,1991. Ontong Java Plateau, Leg 130: synopsis of major drilling results. InKroenke, L.W., Berger, W.H., Janecek, T.R., et al., Proc. ODP, Init. Repts.,130: College Station, TX (Ocean Drilling Program), 497-537.
Berger, W.H., and Wefer, G., 1992. Klimageschichte aus Tiefseesedimenten—Neues vom Ontong Java Plateau (Westpazifik). Naturwissenschaften.,No. 79.
CLIMAP Project Members, 1976. The surface of the ice-age earth. Science,191:1131-1137.
Emiliani, C, 1955. Pleistocene temperatures. J. Geol, 63:538-578.Epstein, S., Buchsbaum, R., Lowenstam, H.A., and Urey, H.C., 1953. Revised
carbonate-water isotopic temperature scale. Geol. Soc. Am. Bull.,64:1315-1325.
Epstein, S., and Mayeda, T., 1953. Variation of O 1 8 content of waters fromnatural sources. Geochim. Cosmochim. Acta, 4:213-224.
Groetsch, J., Wu, G., and Berger, W.H., 1991. Carbonate cycles in the Pacific:reconstruction of saturation fluctuations. In Einsele, G., Ricken, W., Seilacher,A. (Eds.), Cycles and Events in Stratigraphy: Berlin (Springer-Verlag), 110-125.
Hays, J.D., Imbrie, J., and Shackleton, N.J., 1976. Variations in the EartiYsorbit: pacemaker of the ice ages. Science, 194:1121-1132.
Hebbeln, D., Wefer, G., and Berger, W.H., 1990. Pleistocene dissolutionfluctuations from apparent depth of deposition in Core ERDC-127P, west-equatorial Pacific. Mar. Geol., 92:165-176.
Imbrie, J., Hays, J.D., Martinson, D.G., Mclntyre, A., Mix, A.C., Morley, J.J.,Pisias, N.G., Prell, W.L., and Shackleton, N.J., 1984. The orbital theory ofPleistocene climate: support from a revised chronology of the marine δ Orecord. In Berger, A., Imbrie, J., Hays, J., Kukla, G., and Saltzman, B.(Eds.), Milankovitch and Climate (Pt. 1): Understanding the Response toAstronomical Forcing: Dordrecht (D. Reidel), 269-305.
Izett, G., and Obradovich, J., 1991. Dating of the Matuyama-Brunhes bound-ary based on 40Ar/39Ar ages of the Bishop Tuff and Cerro San Luis rhyolite.Geol. Soc. Am. Abstr. Progr., 23:A106.
Johnson, R.G., 1982. Brunhes-Matuyama magnetic reversal dated at 790,000yr B.P. by marine-astronomical correlations. Quat. Res., 17:135-147.
Johnson, T.C., Hamilton, E.L., and Berger, W.H., 1977. Physical properties ofcalcareous ooze: control by dissolution at depth. Mar. Geol, 24:259-277'.
Kroenke, L.W., Berger, W.H., Janecek, T.R., et al., Proc. ODP, Init. Repts.,130: College Station, TX (Ocean Drilling Program).
Le, J., and Shackleton, N.J., 1992. Carbonate dissolution fluctuations in thewestern equatorial Pacific during the late Quaternary. Paleoceanography,7:21-42.
Pisias, N.G., 1976. Late Quaternary sediment of the Panama Basin: sedimen-tation rates, periodicities, and controls of carbonate and opal accumulation.In Cline, R.M., and Hays, J.D. (Eds.), Investigation of Late QuaternaryPaleoceanography and Paleoclimatology. Mem.—Geol. Soc. Am.,145:375-391.
Pisias, N.G., and Moore, T.C., 1981. The evolution of Pleistocene climate: atime series approach. Earth Planet. Sci. Lett., 52:450-458.
Prell, W.L., Imbrie, J., Martinson, D.G., Morley, J.J., Pisias, N.G., Shackleton,N.J., and Streeter, H.F., 1986. Graphic correlation of oxygen stratigraphyapplication to the late Quaternary. Paleoceanography, 1:137-162.
Abbreviations for names of organizations and publication titles in ODPreference lists follow the style given in Chemical Abstracts Service SourceIndex (published by American Chemical Society).
376
QUATERNARY OXYGEN ISOTOPE RECORD, SITE 805
Ramanathan, V., Barkstrom, B.R., and Harrison, E.F., 1989. Climate and theEarth's radiation budget. Physics Today, 42:22-32.
Raymo, M.E., Ruddiman, W.F., Shackleton, N.J., and Oppo, D.W., 1990.Evolution of Atlantic-Pacific δ 1 3C gradients over the last 2.5 m.y. EarthPlanet. Sci. Lett., 97:353-368.
Ruddiman, W.F., Mclntyre, A., and Raymo, M.E., 1986. Matuyama 41,000-year cycles: North Atlantic Ocean and Northern Hemisphere ice sheets.Earth Planet. Sci. Lett., 80:117-129.
Schiffelbein, P., 1984. Stable isotope systematics in Pleistocene deep-seasediment records [Ph.D dissert.]. Univ. of Calif, at San Diego, La Jolla.
Shackleton, N.J., and Opdyke, N.D., 1973. Oxygen isotope and paleomagneticstratigraphy of equatorial Pacific Core V28-238: oxygen isotope tempera-tures and ice volumes on a 105 year and 106 year scale. Quat. Res., 3:39-55.
, 1976. Oxygen-isotope and paleomagnetic stratigraphy of PacificCore V28-239: late Pliocene to latest Pleistocene. In Cline, R.M., andHays, J.D. (Eds.), Investigation of Late Quaternary Paleoceanographyand Paleoclimatology. Mem.—Geol. Soc. Am., 145:449^64.
-, 1977. Oxygen isotope and paleomagnetic evidence for early North-ern Hemisphere glaciation. Nature, 270:216-219.
Shackleton, N.J., Berger, A., and Peltier, W.R., 1990. An alternative astronomi-cal calibration of the lower Pleistocene time scale based on ODP Site 677.Trans. R. Soc. Edinburgh, Earth Sci, 81:251-261.
Shipboard Scientific Party, 1991. Site 805. In Kroenke, L.W., Berger, W.H.,Janecek, T.R., et al., Proc. ODP, Init. Repts., 130: College Station, TX(Ocean Drilling Program), 223-290.
Wefer, G., and Berger, W.H., 1991. Isotope paleontology: growth and compo-sition of extant calcareous species. Mar. Geol., 100:207-248.
Whitman, J., and Berger, W.H., 1992. Pliocene-Pleistocene oxygen isotoperecord, Site 586, Ontong Java Plateau. Mar. Micropaleontol, 18:171-198.
Woodruff, F., Savin, S.M., and Douglas, R.G., 1981. Miocene stable isotoperecord: a detailed deep Pacific Ocean study and its paleoclimatic implica-tions. Science, 212:665-668.
Wu, G., and Berger, W.H., 1989. Planktonic foraminifera: differential dissolu-tion and the Quaternary stable isotope record in the west equatorial Pacific.Paleoceanography, 4:181-198.
, 1991. Pleistocene δ 1 8 θ records from Ontong-Java Plateau: effectsof winnowing and dissolution. Mar. Geol., 96:193-209.
Date of initial receipt: 27 January 1992Date of acceptance: 1 September 1992Ms 130B-032
40 50 60Age (ka)
100
„ 1.2mQ^ 1.0-
p 0.8<o
1 0.6
0.4
0.2
0-5-cycle stack 500-1000 ka
ISR
^_y \^~-SOX
r~—^^\ POX
/Λ r^v—s \ ^J\×~^ DIF
2.0
1.8
1.6
10 20 30 40 50 60Age (ka)
70 80 90 100
0 10 20 30 40 50 60 70 80 90 100Age (ka)
Figure 19. Stack of δ 1 8 θ sections of G. sacculifer (SOX) and Pulleniatina (POX) modulo 100 ka, for five intervals at a time. This "autostack" is meant to capturethe essential trends in the eccentricity domain. A. 0-500 ka. B. 500-1000 ka. C. 1000-1500 ka. ISR = instantaneous sedimentation rate, and DIF = differencebetween SOX and POX.
377
W.H.BERGERETAL.
1.4
~ 1.2mD11 1.0-JO 0.8-•o' 0.6-
0.4-
0.2
OH
ISR
6-cycle stack 0-250 ka
\ ^ ^ ^
DIF
10 15 20 25 30Age (ka)
35 40 45 20 25 30Age (ka)
~ 1.2-
°- 1.0éO 0.8-
1 0.6-
0.4-
0.2
o-
^ ^ ^
6-cycle stack 500-750 ka
ISR
SOX
^ ^ ~
/ ^ POX
' •
DIF
2.2
2.0
10 15 20 25 30Age (ka)
35
-1.8
40 45
CVI
1.0
0.8
0.6
0.4-
0.?
n-
- ^ ^ ^
- ^ — \ _
6-cycle stack 750-1000 ka
ISR
^ ^ ― — - ^ s o x
^ ^ - ^ ^ POX
2.2
2.0
1.8
1.6
10 15 20 25 30Age (ka)
35 40 45
1.4
~ 12mQα 1.0
O 0.8*"O
1 o.6^
0.4
0.2
0
6-cycle stack 1000-1250 ka
0 10 15 20 25Age (ka)
ISR
30 35
SOX
POX
40
2.2
2.0
1.8
45 20 25Age (ka)
Figure 20. Stack of δ 1 8 θ sections of G. sacculifer (SOX) and Pulleniatina (POX) modulo 41 ka, for six consecutive intervals for the time spans shown in
lower left of the graph. A-F. Successively deeper sections in Hole 805C, first three cores.
Figure 21. Correlation between the δ' 8O records of G. saccuUfer in Holes 806B
and 805C with the depths of Site 806 adjusted to those of Site 805, using the
match between the signals. Triangles indicate positions of core breaks. B =
Brunhes, and M = Matuyama.
2.5
2.0
-0.5
G. saccuUfer
o33
DIF (806B-805C)(Fourier fit >86k.y.)
200 400 600 800 1000 1200
Age (ka)1400 1600 200 400 600 800 1000 1200
Age (ka)1400 1600
Figure 22. Comparison of the δ 1 8 θ records of G. saccuUfer in Holes 805C and 806B in the time domain. A. Difference in δ 1 8 θ (DIF), with periods shorter
than eccentricity eliminated. B. Difference in the eccentricity band, illustrating the importance of apparent phase shifts in producing apparent dissolution
cycles (i.e., difference cycles). These shifts may be an artifact of analysis and correlation.
0.6MBDI EBDI EMDI
SOX (805-806)
200 400 600 800 1000 1200 1400 1600Age (ka)
Figure 23. Comparison of difference in δ 1 8 θ values of G. saccuUfer in the
Quaternary records of Holes 805C and 806B (labeled SOX(805-806)) with
difference in δ 1 8 θ values of G. saccuUfer and Pulleniatina in Hole 805C
(labeled 805(SOX-POX)). Dissolution effects on isotopes should produce
upward excursions in either curve. MBDI = mid-Brunhes dissolution interval,
EBDI = early Brunhes dissolution interval, and EMDI = early Matuyama
dissolution interval.
-0.5
-1.0
SRR (806/805)
805 (SOX-POX)
2.0
0 200 400 600 800 1000 1200 1400 1600Age (ka)
Figure 24. Comparison of presumed indices of dissolution SOX (805-806) and
805 (SOX-POX) (as in Fig. 23) with the sedimentation rate ratio between
Holes 806B and 805C. A higher ratio should indicate increased removal from
the deeper site (805C), relative to the shallower one (806B). Ideally, the three
curves should be parallel (if dissolution effects control sedimentation rates at
